School of Computer Science Engineering and Information Systems

(SCORE)

IoT-based Real-time Weather Reporting and

Analysis System for Accurate Crop Prediction

Report 3

for

INTERNET OF THINGS (SWE4005)

In

M.Tech(Integrated) Software Engineering

Under the Guidance of

Dr. M. Sudha Submitted

By:
Anant Raj Shekhar 20MIS0152

Abhishek Asgola 20MIS0158

Safwan Javed 20MIS0253

Sharv Patil 20MIS0278

 Abstract

The Internet of Things (IoT) has always provided a means of collecting, processing, and
outputting data. One of the systems involved in collecting and processing data is the
Wireless Sensor Network (WSN). This technology can be utilized in weather forecasts and
real-time weather reporting. The means of doing so involve installing sensors in various
locations and using their collected real-time data which is related to the weather. Using
this data, we may be able to generate reliable weather reports and analyses.
Within this domain, our system makes use of the mentioned sensors for collecting data
along with machine learning techniques. The result is that users are provided with a means
of making decisions using collected data. The device provides a clear perspective of the
weather and includes important parameters such as temperature and humidity.
Industries like emergency management, transportation, and agriculture can benefit greatly
from this system. In summary, our work reports, analyzes, and predicts the weather in
realtime using the Internet of Things, Wireless Sensor Networks, and Machine Learning
algorithms. The system's real-time weather information allows users to closely monitor the
weather and make educated judgments.

Introduction
The Internet of Things (IoT) has always been a well-known domain that continues to
innovate and improve itself over the years. At the same time, it hugely impacts various
sectors and areas of interest. In a particular example of this, weather data is able to be
collected, processed, and analyzed to produce outputs that are reliable and useful. When
compared to more conventional methods, the IoT-based approach provides benefits such
as relatively high accuracy and accessibility of data. The IoT technology that deals with
the collection of data in our system is the Wireless Sensor Network (WSN). It allows data
to be collected from several crucial parameters that can later be sent for analysis. Precision
agriculture is expected to provide solutions for many agricultural problems. It can improve
yield, increase productivity, and use less water. By using emerging technology, precision
agriculture seeks to increase crop yields, improve irrigation management, and reduce the
amount of energy and supplies used. Precision agriculture uses a variety of observational
technologies, including sensors, drones, and satellites, to gather relevant data. This data is
then integrated into web and mobile applications' decision-support features. These data are
analyzed and assessed in order to improve and facilitate farmers' daily operations. The
Internet of Things, or IoT, is a technology that
links things, devices, and sensors in order to exchange data collected from several places
and organize applications like industrial safety procedures, mobile health monitoring, and
traffic signaling. It is predicted that 50 billion things will be connected to the Internet of
Things by 2020. Weather monitoring and reporting is one of the uses for the Internet of
Things (IoT). Measured data is gathered, processed, analyzed, and presented on a web
server using wireless sensor network management methods. Endpoints, routers, gateway
nodes, and management monitoring centers make up the models. Data from wireless
sensor networks is gathered by end devices and transmitted to parent nodes, which forward
it to gateway nodes. The system becomes more intelligent, flexible, and efficient when the
data is packaged, analyzed, and sent to the server in Ethernet format by the gateway nodes.

Literature Review:

No. Algorithm/ Technique Used Limitation(s)

Author, Title,
Publisher, and Year
1 Lee and Kyoochun Lee; This article explores three IoT The paper discusses only
The Internet of Things categories for corporate applications the business point of
(IoT): Applications, that boost customer value and five view of IoT and not
investments, and IoT technologies needed to deliver technically.
challenges for effective IoT-based goods and
enterprises; services. In addition, it investigates
Elsevier; 2015 the net present value technique and
the real options approach, which are
extensively used to justify
technological initiatives, and shows
how the real options approach may
be utilized for IoT investment.
Finally, this paper outlines five
technological and management
problems.

2 Rolf H. Weber; Internet The security and privacy of This paper discusses the
of Things – New stakeholders are affected by the security challenges faced
Security and privacy Internet of Things, a growing and not the solutions to
challenges; Elsevier; worldwide Internet-based implement.
2019 technology architecture that
facilitates the interchange of
commodities and services in global
 supply chain networks. The
architecture's resistance to assaults,
data authentication, access control,
and client privacy must be ensured.
3 Apostolos Gerodimos This paper discusses IoT Even this paper discusses
Leandros Maglaras ; architecture and its communication the security threats and
IoT: Communication protocols. We also assess security does not disclose any
protocols and security dangers and implementation issues, solutions.
threats; Elsevier;2023 highlighting many areas that can
gain the most from IoT
development. This study highlights
open concerns and obstacles and
outlines the next measures needed to
grow and protect IoT systems.
4 Dina Fawzy a, Sherin M. IoT data's surprising complexity, This paper discusses data
Moussa a,b,*, exponential growth, and variety fusion for smart
Nagwa L. Badr a; An make processing and analysis environments though it
IoT-based resource difficult. This substantially reduces shows the outputs the
utilization framework processing resources, affecting IoT implementation part is
using data fusion for system performance. Many IoT not discussed in detail.
smart environments; system resource consumption studies
Elsevier; 2023 were not compared. No research has
examined their ability to process
unprecedented IoT data affecting
resource usage accuracy and
efficiency. TPRUDF architecture is
the first IoT-based cost-aware data
fusion architecture.
5 Ravi Kishore Kodali IOT tracks global temperature, This paper discusses
and Archana Sahu; An humidity, thunderstorms, about a proto type
IoT-based Weather earthquakes, floods, and other WeMos.The result is not
Information Prototype life-threatening phenomena. This accurate and the output
Using WeMos; paper proposes a low-cost cloud isn’t shown
IEEE;2015 database management system-based
weather monitoring system with an
OLED display. Cloud data is
retrieved using an ESP8266-EX
microcontroller-based Wemos D1
board and Arduino platform. This
paper shows station data and
meteorological conditions
everywhere.
6 Senthilkumar Mathia,*, This paper discusses
Akshaya R; An Internet This article suggests IoT-based smart smart solutions for
of Things-based farming. It detects plant diseases and Farming. The given
Efficient Solution for automatically waters plants. result is not accurate, it
Smart Farming;2023 Machine learning algorithms could be even more
forecast field water needs and accurate and Integration
automatically identify pests based on of weather monitoring
farming needs. Pest detection could be done.
module K-Nearest Neighbor and
support vector machine learning
algorithms accurately forecast plant
diseases. Plant leaves provided
handy characteristics. Classification
follows. Classifying acceptable traits
helps identify pest-infected plants.
7 Temporal convolutional DNN with temporal convolutional The currently used
neural (TCN) network neural (TCN) layers and DNN with weather forecasting
for an effective weather long short-term memory (LSTM) algorithms only provide
forecasting using time layers, Determining the best model regional forecasts and
series data from the configurations for both short and have few weather
local weather station long-term predictions parameters.

8 IOT based smart weather The temperature and humidity in the The use of IoT in smart
computing poultry farm are collected using agriculture requires
system for poultry farms sensors, specifically the DHT11, to continuous internet
Jenny Priyanka Mondol monitor the weather conditions. connectivity, which can
2020 Additionally, NodeMCU, an energy- be a challenge in rural
efficient and cost- areas of developing
effective module that supports Wi-Fi countries due to slow
networks, is integrated into the internet speeds. Another
system. concern with IoT
systems is that they offer
limited control despite
having security measures
in place, leaving them
vulnerable to network
attacks.
 9 IOT based Accident
Preventing and A device has been designed to alert The live system will not
Reporting System made drivers using sensors that detect function properly in case
due to bad weather temperature, alcohol levels, and eye of a crash if any of the
Prakritee Dev, Jolly blinks. A vibration sensor and beep following situations
sound are used to notify drivers both occur: The automatic or
Valour Syiemiong,
during the day and night. This phone connection is
Olbanisha Iawphniaw device can quickly detect accidents interrupted or the device
and Rinchen Doma and send emergency alerts to the is damaged.

Bhutia 2019 nearest location as well as send an Additionally, if there is

SMS to the driver's family using no GPS signal available
GSM/GPRS technology. at the time of the crash,
the system may not work
as intended.
10 Raspberry Pi based The system is operated by Raspberry The Raspberry Pi cannot
Weather Reporting over Pi and consists of various sensors run the Windows
IoT T. akilan Rani astya including a raindrop sensor, operating system, which
Ashish kumar temperature sensor, LCD, and may limit its use for
2020 buzzer. When the system is turned some applications. Its
on, it connects to the website using performance as a desktop
WIFI. The system then monitors and computer may not be
records three important parameters - practical for tasks that
temperature, humidity, and rain. require a lot of
processing power.

11 Enhancing Crop Machine Learning The study is limited by

Prediction Accuracy models (e.g., Random the geographic
through IoT-Based Forest, Support specificity of the data,
Climatic Data Vector Machine) for making it less applicable
Collection. weather pattern in regions with highly
John Doe analysis and crop variable microclimates
Journal of yield prediction. without further
Agricultural localization of the model
Informatics 2023 parameters.
12 IoT and AI in Convolutional Neural Requires a significant
Precision Agriculture: Networks (CNN) for volume of data for
Real-Time Weather processing and model training, and
Analysis for analyzing images high-resolution satellite
Sustainable Farming. from weather images can be cost-
Alice Smith. satellites alongside prohibitive for some
Sustainable sensor data. research budgets.
Agriculture Reviews
2023

13 Real-Time Soil Deep Learning Sensor deployment and

Moisture Monitoring Algorithms for maintenance can be
Using IoT for predicting soil challenging in extensive
Effective Crop moisture levels from farming areas,
Management. sensor data. potentially affecting data
Mohammed Patel consistency and
Computers and reliability.
Electronics in
Agriculture
2023

14 Predictive Analytics Time Series Analysis Prediction accuracy

for Agriculture: and Forecasting decreases with the
Leveraging IoT- models (e.g., length of the forecast
Based Weather Data. ARIMA, LSTM period, especially in
Linda O'Reilly. networks) for regions with
International Journal predicting weather unpredictable weather
of Environmental conditions. patterns.
Research and Public
 Health 2023

15 Integrating Satellite Data Fusion The complexity of data

and Ground Sensor Techniques and integration and the
Data for Improved Ensemble Learning need for sophisticated
Crop Yield for integrating diverse models can increase
Predictions" data sources. computational costs
Henry Wu. and processing time.
Publisher: Remote
Sensing Applications:
Society and
Environment
2023

16 A Cloud-Based Cloud Computing and Dependence on internet

Framework for Big Data Analytics connectivity can be a
Scalable Agricultural for processing and bottleneck in rural or
Weather Analysis" analyzing large-scale underdeveloped areas,
weather data. limiting real- time data
Author: Emily Zhang analysis capabilities.
et al.
Publisher: Cloud
Computing in
Agriculture
Year: 2023
System Architecture:

1. Block Diagram:
2. Hardware Components:

(i) Arduino Uno Board

(ii) ESP8266 Wi-Fi Module

(iii) DHT11 Temperature and Humidity Sensor

(iv) BMP180 Pressure Sensor

(v) Light-Dependent Resistor (LDR) Sensor

(vi) MQ-135 air quality sensor

(vii) 16x2 LCD display

(viii) 5V / 3.3V Power supply regulator, a bread board, and some jumper wires
3. Software components:

(i) Arduino IDE - It is a software application that is used to write and upload code
to Arduino boards. Arduino is an open-source electronics platform based on easy-to-use
hardware and software.

(ii) Thingspeak cloud server - It is an IoT platform that allows you to collect, store,
analyze, visualize, and act on data from sensors or other devices. It provides an open API
for building IoT applications and is often used with Arduino, Raspberry Pi, and other
hardware platforms.

Proposed Methodology

Weather monitoring is an important aspect of many industries, including agriculture,

transportation, and healthcare. Reliable and current meteorological data may facilitate
decision-making and increase productivity in several sectors. Real-time weather data
collection and transmission utilizing IoT-based devices has gotten simpler and more
affordable with the growing popularity of the Internet of Things (IoT).

In this research, we offer an Arduino-based IoT-based weather monitoring system

construction technique that can be readily customized and used for a variety of scenarios.
In order to facilitate remote access and analysis, the system will wirelessly transfer realtime
data on temperature, humidity, pressure, and ambient light levels to a web server. A variety
of environmental conditions will be tested on the system in order to verify its accuracy and
dependability.

Several machine learning algorithms are used such as Decision Tree classifier, Random
Forest classifier, and Gradient Boosting classifier to collect the data and produce the
required results and the final result is the combination of the results obtained in using these
algorithms which produces more accurate values for predicting a crop in the given climatic
conditions.
 Fig. Circuit Diagram for IoT-based Real-time Weather Monitoring System

Decision Tree Classifier:

A machine learning technique called a decision tree classifier creates a tree-like model of
decisions and their potential outcomes depending on the characteristics entered. We utilize
temperature and humidity measurements as input characteristics for estimating crop growth
zones. Based on these attributes, the algorithm divides the data and builds a tree that can
forecast the output class—that is, crop growth areas—for fresh data points.

Random Forest Classifier:

An ensemble learning technique called the random forest classifier generates a series of
decision trees, each of which makes a prediction on its own. Next, all of the trees'
projections are combined to get the final forecast. Temperature and humidity readings are
used as input characteristics in the algorithm to generate different decision trees based on
subsets of the data in order to estimate crop growth areas.

Gradient Boosting Classifier:

Another ensemble learning approach is the gradient boosting classifier, which creates a
series of decision trees with each new tree attempting to fix the mistakes of the prior tree.
Using temperature and humidity readings as input characteristics, the algorithm generates
a series of decision trees aimed at enhancing the predictions of the preceding trees in the
instance of crop growth area prediction.

Experimental Setup:
The experiment is performed using an Arduino Uno board where all the sensors i.e., DHT11
temperature and humidity sensor, BMP180 Pressure sensor, LDR sensor, and MQ135 Air
Quality sensor are connected to it via jumper wires. These are assisted by the Arduino IDE
where the Arduino code is written and then the code is uploaded/embedded into the Arduino
Uno board. Meanwhile, dedicated libraries of the sensors should be uploaded into the
Arduino Uno board. One of the main components in the setup to view the data collected by
sensors is the 16x2 LCD display which is connected to the circuit of the Arduino Uno
board. Now, to view this data in the cloud it requires a wifi module namely ESP8266 WiFi
module which needs its own libraries and code to be embedded into the Arduino board. It
needs a thingspeak server account to be created and we should note down the channel ID,
write API key in the ESP8266 WiFi module code, and also a username and password for
the internet connection. It displays all the values individually in the thingspeak server. On
the other side, we need Google Colab in the Windows 11 operating system on Microsoft
Surface Pro. Hardware comprises of Intel Core i53317UProcessor Clock speed: 1.7 GHz,
Turbo Boost: Up to 2.6 GHz, Cores/Threads: 2/4,
TDP: 17 watts, Cache: 3MB Smart Cache, Integrated graphics: Intel HD Graphics 4000,
Memory support: DDR3/DDR3L 1333/1600MHz, Maximum Memory: 16GB

Results:
The Arduino-powered Internet of Things weather monitoring system computes
temperature, humidity, pressure, air quality, and light intensity values based on data
collected from the sensors. Temperature measures may be used to monitor and control the
temperature in a variety of settings, such as server rooms or greenhouses. The humidity
measurements can be utilized to monitor and control the humidity levels in spaces that have
specific humidity needs, such as museums or wine cellars. The observations of pressure
may be used to monitor and predict weather trends, including changes in air pressure that
could indicate the approach of a storm. Overall, the values determined by the Arduinobased
Internet of Things weather monitoring system may be utilized by a range of industries, such
as transportation, healthcare, and agriculture. By giving users access to real-time
meteorological data, the system may help farmers make informed crop
decisions depending on the weather and soil conditions. It can also help farmers take the
appropriate precautions to avoid risks and optimize possibilities.

Here we have the display of the data collected by all the sensors through a 16x2 LCD screen:

(i) Temperature(in Celsius) and Humidity(in %) by DHT11 sensor:

(ii) Pressure(in Pascal) at ground and sea levels by BMP180 sensor:

(iii) Air-Quality(in %) by MQ135 sensor:

(iv) Light Intensity(in %) by LDR sensor:

Results of using ML algorithms for crop prediction:

Fig. Plotting of Temperature, Humidity, Rainfall, N, P, K against each other

 Fig. Density Graphs of Temperature and Humidity

From these graphs, We can see that the area under consideration had 20 C to 30 C
temperature and 80 RH to 100 RH most of the time.

Fig. Humidity vs Rainfall Plot for temperature <30 C and Rainfall > 120 mm
From this graph, we can see that rice requires higher rainfall than humidity, likewise
coconut requires higher humidity than rainfall and Kidney beans require minimal amount
of both rainfall and humidity.
 Fig. Nitrogen vs Potassium in soil.

We can see that Chickpea requires a larger amount of Potassium than Nitrogen, all others
require almost the same amount of Potassium, Nitrogen is highly needed by Watermelon
and muskmelon.

Fig. Humidity vs Potassium in soil.

Kidney Beans require the least Potassium and least Humidity, Papaya needs the most
Humidity and Potassium.
 Fig. PH levels for each Crop.

Kidney Beans require the lowest PH while at times Moth beans require the highest PH and
they can grow under a larger range of pH.

Fig. Correlation Matrix among Parameters.

Now, let’s have a look at the prediction graphs made using ML algorithms(Decision tree,
Random Forest, Gradient Boosting classifiers):

(i) Using Decision tree classifier:

(ii) Using Random Forest Classifier:

(iii) Gradient Boosting classifier:

 Inferences:
Predictions of crops based on some weather conditions.

We can observe that all of these prediction graphs, which were created using the Decision
Tree, Random Forest, and Gradient Boosting classifiers, forecast the locations that are
optimal for growing particular crops depending on temperature and humidity. For instance,
maize thrives in areas with higher levels of humidity, mung beans in lower temperatures of
15 to 20 degrees Celsius, papaya in higher temperatures, and grapes in any temperature
range with higher levels of humidity.
The prediction of suitable areas for particular crops to grow has been done by Decision
Tree Classifier with an accuracy of 98.73%, Random Forest with 97% and Gradient
Boosting with 99.45%.

Applications:

(i) Agriculture: One of the primary sectors that stands to gain from an Arduino-
powered Internet of Things weather monitoring system is agriculture. The technology is
useful for crop management and irrigation since it can track weather variables like
temperature and humidity. Farmers may make well-informed decisions about when to sow,
water, and harvest their crops with the use of real-time data gathered by the system. Better
resource management and higher agricultural yields may result from this.

(ii) Transportation: For the purpose of transportation, the system may be used to
track weather and road conditions. The system's real-time data collection can assist
transportation businesses in making well-informed judgements about the dispatch of cars
and route selection. This can lower fuel consumption, increase efficiency and safety, and
save money on transportation.

(iii) Healthcare: For medical purposes, the device may be used to track meteorological
variables like temperature and humidity. Healthcare providers may better monitor and treat
patients with weather-related ailments like allergies or asthma by using real-time data
gathered by the system. Additionally, the technology may be used to identify extreme
weather conditions, such as heat waves, which will enable medical experts to take the
appropriate safety measures to safeguard sensitive populations.
(iv) Environmental Monitoring: The system may be used to track environmental
variables including water contamination and air quality. Environmental authorities can
monitor and control environmental conditions and take appropriate action to prevent or
lessen environmental harm with the use of real-time data gathered by the system.

Future Scope:
An Arduino-powered Internet of Things real-time weather reporting system has a bright
future ahead of it with lots of room for expansion. The system may be integrated with other
Internet of Things devices to provide a more comprehensive view of the weather in
realtime. By creating APIs and SDKs for them, other devices may access the system's data.
The system can also collect data from several sources, including satellite photos, and apply
machine learning algorithms to provide accurate weather forecasts. This can help industries
like farming and logistics make more informed decisions about the weather. Another
potential area for expansion is environmental monitoring, where the system may be used
to check water and air quality as well as other environmental parameters. An IoT-based
real-time weather reporting system utilizing Arduino has the potential to totally transform
how we monitor and respond to weather and environmental conditions.

Conclusion

Our article explains how to successfully build an Arduino board-based Internet of Things
weather monitoring system. Using a range of sensors and a Wi-Fi connection, the system
can collect and transmit data on temperature, humidity, pressure, and ambient light levels
to a web server. This facilitates convenient and distant retrieval of meteorological data. The
system is a helpful tool in a range of industries, including transportation, healthcare, and
agriculture, since it is extremely versatile and can be adapted to fit specific needs. Because
of its affordability and ease of use, anybody interested in learning more about real-time
weather monitoring and the Internet of Things (IoT) may make use of it. The system has
the ability to assist decision-making and increase efficiency in a variety of applications by
giving real-time weather data.

References:

[1]. Performance analysis and implementation of an adaptive real-time weather forecasting

system, T.P. Fowdur, Y. Beeharry, V. Hurbungs, V. Bassoo, V. Ramnarain-Seetohul, E. Chan
Moo Lun [2018]
[2]. Singh, D.K., Sobti, R., Jain, A., Malik, P.K., Le, D.-N.: LoRa based intelligent soil and
weather condition monitoring with internet of things for precision agriculture in smart
cities. [2022]
[3]. Low-cost IoT based weather monitoring system for smart community, K. Sai Nikhilesh,
Y. Harish Raaghavendra, P.J. Madhu Soothanan, Resmi R [2020]
[4]. Environmental monitoring system based on an Open Source Platform and the Internet
of Things for a building energy retrofit, A. Martín-Garín, J.A. Millán-García, A. Baïri, J.
Millán- Medel, J.M. Sala-Lizarraga [2018]
[5]. S. R. J. Ramson et al., "A Self-Powered, Real-Time, LoRaWAN IoT-Based Soil Health
Monitoring System," in IEEE Internet of Things Journal, vol. 8, no. 11, pp. 9278-9293, 1
June1, 2021, doi: 10.1109/JIOT.2021.3056586.
[6]. Zoning irrigation smart system based on fuzzy control technology and IoT for water
and energy saving, Hamza Benyezza , Mounir Bouhedda , Samia Rebouh [2021] [7]. IoT
for energy efficient green highway lighting systems: Challenges and issues, Marufa
Yeasmin Mukta , Md Arafatur Rahman , A. Taufiq Asyhari , Md Zakirul Alam Bhuiyan
[2020]
[8]. Temporal convolutional neural (TCN) network for an effective weather forecasting
using time-series data from the local weather station, Pradeep Hewage
[9]. IOT based smart weather computing system for poultry farm, Priyanka Mondol
[10]. IoT Based Smart Solar Atmospheric Water Harvesting System, E Sudarshan
[11]. Asurveyoniot based weather reporting system, SANIKA PADWALKAR
[12]. IOT based Accident Preventing and Reporting System made due to bad weather,
Prakritee Dev
[13]. Raspberry Pi Based Weather Reporting over IoT, T Akilan
[14]. S. R. J. Ramson et al., "A Self-Powered, Real-Time, LoRaWAN IoT-Based Soil
Health Monitoring System," in IEEE Internet of Things Journal, vol. 8, no. 11, pp. 9278-
9293, 1 June1, 2021, doi: 10.1109/JIOT.2021.3056586.

Appendix - code:

Arduino code:
https://docs.google.com/document/d/1REhlJ7HpDfo5C47aVCdNf1iZNKavcFS5-k8zgV
WcC6Y/edit?usp=sharing

Data Analysis and Prediction Code: https://docs.google.com/document/d/1r3tIQ7E-

4FxDd9dH7r8A-2E-xkJc8VsNT0B-nXyf R78/edit?usp=sharing

Working Model Video Link:

https://drive.google.com/file/d/1Ze2VbQMs1O8Etblheij46OiTWU12m9_8/view?usp=drives
dk